A two-stage waveform-driven cross-differential protection scheme for PV-integrated parallel transmission lines

The increasing penetration of photovoltaic (PV) generation in modern power systems introduces significant challenges to the protection of parallel transmission lines, primarily due to limited and highly variable fault current characteristics. Conventional cross-differential relay (CDR) schemes provide fast and selective protection for parallel transmission lines, but their reliance on current-magnitude comparison can limit their discrimination capability in PV-integrated systems. An enhanced CDR approach has addressed this limitation by incorporating remote-end voltage information to estimate the grid-voltage angle. However, this estimation relies on voltage measurement at the PV bus, the line impedance, and the equivalent grid impedance determined using voltage and current measurements at the grid bus. To address these measurement requirements, this paper proposes a two-stage waveform-driven cross-differential protection scheme for PV-integrated parallel transmission lines that relies solely on the currents of the parallel lines measured at the PV bus. In the first stage, internal fault detection is achieved using a normalized cross-correlation (NCC)-based fault detection index that captures waveform similarity between parallel line currents, thereby reducing the dependence on current amplitude under challenging operating conditions. In the second stage, faulty line identification is performed using the polarity of the fault-imposed positive-sequence current extracted via a delta filtering approach, enabling reliable direction-sensitive decision-making without requiring voltage measurements or communication links. The proposed method is extensively evaluated on a modified IEEE 39-bus system under diverse scenarios, including high-resistance faults, measurement noise, line parameter asymmetry, varying fault inception angles, PV generation intermittency, grid-code compliant reactive power support, and external faults. Simulation results demonstrate that the proposed scheme achieves fast, robust, and accurate fault detection and identification across a wide range of operating conditions.